tech-kern archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: kernel goes dark on boot



Le 16/01/2023 à 10:38, Emmanuel Dreyfus a écrit :
On Mon, Jan 16, 2023 at 09:08:20AM +0000, Emmanuel Dreyfus wrote:
On Mon, Jan 16, 2023 at 06:50:18AM +0000, Emmanuel Dreyfus wrote:
Clearly broken firmware - is there an update available?
Perhaps I can have more luck booting through grub, I will
try that.

Same problem, grub works fine, and display goes dark as soon as
the kernel is supposed to display something.

I also tried NetBSD's netive EFI bootstrap with Xen and latest
netbsd-XEN3_DOM0. Xen displays boot messages until
xen is relinquishing vga console (not sure about the last two
words, it goes really fast), then a dark screen.


Upping the thread, I may experience a similar situation, related to this code. At least it shows a "lookalike" behavior during boot, which I have a hard time tracking down.

It happens with NetBSD-10 on a U59 (a NUC derivative based on a celeron N5105).

1 => when booting with a screen connected to its HDMI port, all goes well; the Intel GFX card (supported in our i915 drm code btw) prints out all the expected console output, and the efiboot => kernel handover goes fine;

2 => when I go headless with that system (nothing plugged to HDMIs), just after efiboot transfer to the kernel, it seems to crash without having any console output, screen gets blank.

Whether efiboot or kernel is faulty is hard to tell (I lose screen and have no serial available on that thing); the culprit seems to happen in the same part of the kernel boot though, eg. early. If I unplug the HDMI during the memory mappings loading (spinning wheel with all the numbers, the usual last prints from boot(8)), the boot will fail. The only way out is a hard reset, it does not seem to fall in ddb(4) either. Typing things blindly does not yield any result.

This is related to graphics handling at early boot it seems, the only difference between both is having a screen connected or not in my case. I did the same tests manu@ suggested (grub especially), I can reproduce the issue in the same way.

Methodology in narrowing this one down appreciated. Thanks :)

Cheers,
jym@
--
Jean-Yves Migeon
The NetBSD Foundation
http://www.NetBSD.org



Home | Main Index | Thread Index | Old Index